Chaofang Hu

In general, a rotating reference frame moving along the reference path is established in the unmanned surface vehicle (USV) path following method based on line-of-sight (LOS) guidance, and good following performance is achieved for the paths with small curvature. However, this method often results in significant deviations at multiple sharp turns due to underactuation. Accordingly, a new path following method via the rotating reference frame is proposed in this article. First, it is proven that there must exists an optimal rotating frame angle, by which USV can follow the reference path precisely. Second, a novel rotating reference frame adaptive law is designed to track this optimal angle, by which the frame angle rotation rate can be regulated for adapting to variation of path curvature and environmental disturbances. Stability analysis proves that the proposed method can follow multisharp-turn reference path in presence of disturbances. In addition, the desired constrained conditions determining the bounds of desired speed and reference path curvature under thruster saturation are developed. Path following can be achieved only if the desired speed and path curvature remain within these bounds. Finally, both simulation and experiment results validate effectiveness of the proposed method by comparing with the existing LOS guidance methods.

For multi-unmanned surface vessel (Multi-USV) capturing target, the coordination control method of path planning and tracking is proposed to integrate multi-agent reinforcement learning (MARL) and active disturbance rejection control (ADRC) dynamically and effectively. The bounded water environment model with various obstacles is constructed. To generate the optimal capturing path online and accelerate its convergence, the real-time multi-agent deep deterministic policy gradient (MADDPG) is enhanced by combining prioritized experience replay (PER). In order to achieve interaction between agent and environment, the real positions of unmanned surface vessels (USVs) are input to the MADDPG network as state variables. The action space consists of yaw angle and surge speed of USV and is reckoned as the reference path of tracking control. In case of target escaping, it is required that USVs are evenly distributed in the target-centered capture loop but not located within the detection range of target. For fast and safe capture, the composite reward function is proposed by designing capture reward, obstacle avoidance and collision avoidance reward, boundary collision restriction reward, capture inner boundary constraint reward, angle constraint reward and motion constraint reward. In addition, to integrate the actual tracking performance, the errors between the references and real states of USVs are also formulated into the reward function. In order to follow the reference commands from the enhanced MADDPG in presence of disturbances of wind and wave, the angle and speed tracking controllers are developed using linear ADRC (LADRC). Finally, the effectiveness of the proposed method is verified by capture simulations of static target and dynamic target with various types of obstacles.

This paper investigates a distributed model predictive control formation method for unmanned aerial vehicles (UAVs) to cope with potential communication delay (CD) and packet dropout (PD) in communication channels during multi-round formation missions. For the impact of historical communication data during execution of each formation mission, these two communication abnormities are modeled as memory-based double random processes (DRPs). Delay sequences and dropout sequences of DRPs that mark the occurrence of CD and PD according to the delay and dropout rates can be generated. Accordingly, the impact of historical data is reckoned as probabilistic increments to be incorporated into formation control. Based on the consecutive occurrence number of delay and dropout, joint occurrence variable is defined. Transmission protocols of the assumed prediction states are formulated for distributed model predictive control. In order to ensure fast convergence of system with abnormal communication states, the effect of memory-based CD and PD is piled into the Lyapunov function and the terminal feedback controller is customized by linear error modeling and linear inequality matrix. Optimization models including different objective functions and constraints are developed for the leader and followers respectively. The system stability using the proposed control method is guaranteed. Finally, the simulation results show that the proposed formation control strategy can effectively mitigate the memory-based communication defects.

This article investigates the formation problem of unmanned aerial vehicles (UAVs) in presence of communication anomalies including delay and data loss. A formation controller based on distributed model predictive control (DMPC) is proposed for the three-dimensional (3D) kinematic model of UAV. According to the different roles of UAVs, asynchronous and synchronous communication methods are used for information interaction in leader-follower (LF) formation. In consideration of communication anomalies, the assumed prediction states are applied to maintain UAVs in the predefined formation, and the updated rules for the assumed states are designed. The formation cost functions are developed for the leader and follower, respectively. The terminal feedback controllers of leader and followers are designed by 3D error model and linear matrix inequality (LMI) to derive the assumed terminal state converging to the desired value in case of communication anomalies. Utilizing the assumed prediction states of the leader and follower in the presence of communication anomalies, the collision avoidance constraint and error constraint are built to ensure the safety and formation accuracy. It is proved that the tracking and formation errors of UAVs can converge to zero in abnormal communication. Finally, simulations are conducted to verify the effectiveness of the proposed method.

In this paper, a novel online tube model predictive control (Tube-MPC) method is proposed for flexible hypersonic vehicle with time-varying input delay. The polytopic linear parameter varying (LPV) model with time-varying input delay is established and further transformed into a polytopic LPV delay-free model by input delay model transformation (IDMT). Since the flexible states cannot be measured accurately due to vehicle body deformation, the polytopic Luenberger observer is developed. To reduce control conservatism, substitute observations are used to replace the actually observed states such that the new reduced control error invariant set (CEI) is generated. On the basis, the Tube-MPC controller is designed, where the offline robust feedback control law and the online nominal control law are devised via linear matrix inequality (LMI) separately. For figuring out the invariant set of nominal control, an offline-online synthesized method is proposed by combining interpolation-based offline computation (IOFC) with substitute observation-based online computation (SOOC). In IOFC, the sequences of observation error invariant set (OEI) and CEI are computed offline iteratively and further interpolated so as to reduce online computation burden. The online judgement and update mechanism is proposed in SOOC. When the observed states enter the expanded invariant set constructed by the interpolated OEI and CEI, the substitute observation-based control error invariant set (SOCEI) will be calculated online and accordingly the nominal control law can also be updated. The simulation results verify the proposed control method’s effectiveness for flexible hypersonic vehicle with time-varying input delay.